Migrations as changesets
Propose schema changes as multi-statement migrations targeting one database — not ad-hoc SQL run straight against production.
Database change management
Vak Shipyard
Ship schema changes the safe way. Propose migrations, validate them against engine-aware rules, get them approved, and apply through a controlled write connection — now or on a schedule — across 22 database engines.
- Pre-flight validation
- 22 database engines
- Approval-gated writes
🔒 shipyard.yourcompany.com/migrations
⛁
Approved
add-fulfilled-at-to-orders
production · orders_db · PostgreSQL
✓ Draft
✓ Pending
3 Approved
4 Applied
ALTER TABLE orders
ADD COLUMN fulfilled_at TIMESTAMP
ADD INDEX idx_orders_status (status);
Reviewers
2 / 3 approved
AS RK MP
RK Ravi K. 2m ago
Index looks good. Safe to apply on the write connection — backfill is nullable. ✅
Reviewed migrations
No more cowboy SQL.
Every schema change is a reviewable migration — validated by engine-aware rules, approved by the right people, and applied on your terms before it ever touches a database.
Pre-flight validation
Engine-aware validation rules check every statement before review — risky patterns are flagged early, with toggles and thresholds you control per engine.
Review & approve
Add reviewers, thread comments, and require as many distinct approvals as your project demands before a single statement runs.
Apply now — or on schedule
Designated releasers apply approved migrations through a dedicated write connection, immediately or scheduled for a quiet window.
🔒 shipyard.yourcompany.com/schema
orders PostgreSQL
Pull schema
id PK
bigint
customer_id FK
bigint
status
varchar(24)
total
numeric(12,2)
fulfilled_at
timestamp
created_at
timestamp
SELECT status, count(*) FROM orders GROUP BY status;
Read-only connection · 4 rows · 38 ms
Browse & query
See your schema, query it safely.
Explore tables and indexes, pull the live structure over a read connection, and run SELECT-only queries in Query Studio — across 22 engines, from PostgreSQL and MySQL to Snowflake, BigQuery and MongoDB.
Schema browser
Explore tables, columns, types, indexes and row estimates across every connected database.
Pull schema
Sync the current structure straight from the database over its read connection — always up to date.
Query Studio
Run SELECT-only queries with a table view or MySQL \G-style vertical view — safe by construction.
Saved & shared queries
Save queries with names, tags and descriptions, and share them with the team via links — your tribal SQL, kept in one place.
22 engines
One workflow, every database.
Relational
- PostgreSQL
- Aurora PostgreSQL
- AlloyDB
- MySQL
- Aurora MySQL
- MariaDB
- TiDB
- Oracle
- SQL Server
Analytics & warehouse
- ClickHouse
- Snowflake
- BigQuery
- Redshift
- Databricks
- StarRocks
- Elasticsearch
NoSQL
- MongoDB
- Redis
- Cassandra
- DocumentDB
- DynamoDB
- Cosmos DB
See it govern your database changes.
A short call is the fastest way to see Vak Shipyard on your stack.
Team & access
Built for a whole team.
Roles, an in-app Users console, read/write connection guardrails, and a full audit log — give your team database access without handing over prod.
A dashboard that assigns work
Everyone gets a needs-your-action feed — approve, release, review or fix — plus activity trends and per-database change counts.
Per-project governance
Each project sets its own approvers, releasers, required approval count and self-approval policy — the rules live with the project, not a wiki.
Connection guardrails
Separate read and write connections per database mean schema pulls and queries can never mutate data.
System-wide audit log
Every action — migrations, approvals, users, settings — is recorded in a filterable, admin-only audit log.
Own your change process
Self-host it, keep control.
Schema changes touch your most critical data. Vak Shipyard runs in your own environment, with every write reviewed and recorded.
Self-host anywhere
Run it on your own infrastructure so your schemas, credentials and migration history never leave your control.
Approval-gated writes
Nothing touches production until it's reviewed and approved — the write path is governed end to end.
Environment-aware
Production, staging and development are first-class, so a change is always scoped to the right place.
Email & Slack alerts
SMTP email plus per-event Slack notifications — on submit, approve, apply, and reviewer assignment — so nothing waits unseen.
Pricing
Own it today. Let us run it later.
Start with the complete one-time package. Special features, integrations, and bespoke workflows can be quoted separately.
Lifetime deal · first 100 users only
One-time
Available now $199 $499
Whole package. Buy once, self-hosted, yours forever.
- Lifetime license, yours to keep
- Self-host on your infrastructure
- Unlimited projects & databases
- Onboarding & email support
Custom features
Add-onsCustom
For bespoke workflows, special integrations, and engine support.
- New database engine support
- Custom approval workflows
- SSO, roles, or data model changes
- Dedicated implementation support
Managed
Available in beta
$10 / month
We host, update and maintain Shipyard.
- Everything in One-time
- We deploy, host & scale it
- Monitoring & auto-updates
- Priority support & SLAs
FAQ
Questions, answered.
What is Vak Shipyard? +
A self-hostable database migration and schema-management tool. Propose schema changes as reviewable migrations, get them approved, apply them through a controlled write connection, and keep a full audit trail — all behind sign-in with role-based access.
How does the migration workflow work? +
Each migration targets one database and carries one or more ordered SQL statements. It is validated against engine-aware rules, then moves through a lifecycle — draft, pending approval, approved, applied (or rejected / failed) — with reviewers and threaded comments along the way. Only approved migrations are applied, always through a dedicated write connection, immediately or scheduled for a later time.
Which databases does it support? +
22 engines across relational, analytics and NoSQL systems: PostgreSQL, Aurora PostgreSQL, AlloyDB, MySQL, Aurora MySQL, MariaDB, TiDB, Oracle, SQL Server, ClickHouse, Snowflake, BigQuery, Redshift, Databricks, StarRocks, Elasticsearch, MongoDB, Redis, Cassandra, DocumentDB, DynamoDB and Cosmos DB — organized as Project → Environment → Database. Capabilities are gated per engine, and more engines can be added for enterprise deployments.
What are validation rules? +
Pre-flight checks that run against a migration's SQL before it is submitted for review — for example flagging destructive statements, missing indexes or risky patterns. Rules are engine-aware, can be toggled on or off, and support configurable thresholds, so admins decide how strict each engine's checks should be.
Can approval rules differ by project? +
Yes. Each project defines its own governance: who may approve, who may release (apply) approved migrations, how many distinct approvals are required, and whether authors may approve their own migrations.
Can people query data without risk of changing it? +
Yes. Each database has separate read and write connections. Schema pulls and Query Studio use the read-only connection and accept SELECT-only queries, so browsing and querying can never mutate your data. Useful queries can be saved with tags and shared with the team via links. Writes only happen via approved migrations.
Can my whole team use it, with different permissions? +
Yes. Users have admin, editor or viewer roles managed from an in-app Users console. Viewers browse and query; editors create, submit and comment on migrations; admins approve, apply, and manage users, connections and settings. Every action is recorded in a system-wide audit log.
Can I self-host, or do you run it for me? +
Both. Buy the one-time license and run it on your own cloud so schemas and credentials never leave your control, or choose managed deployment and we host, scale and maintain it for you.
